đ swin-tiny-patch4-window7-224-finetuned-new_dataset_50e
This model is a fine - tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It offers significant value in image classification tasks by leveraging the pre - trained architecture and fine - tuning on specific data, achieving high accuracy on the evaluation set.
đ Documentation
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
Training Loss |
Epoch |
Step |
Validation Loss |
Accuracy |
No log |
0.94 |
4 |
0.7081 |
0.6081 |
No log |
1.94 |
8 |
0.7104 |
0.6351 |
0.5516 |
2.94 |
12 |
0.6911 |
0.6351 |
0.5516 |
3.94 |
16 |
0.7156 |
0.7027 |
0.537 |
4.94 |
20 |
0.7345 |
0.7297 |
0.537 |
5.94 |
24 |
0.6745 |
0.6892 |
0.537 |
6.94 |
28 |
0.7146 |
0.7297 |
0.5333 |
7.94 |
32 |
0.7057 |
0.6892 |
0.5333 |
8.94 |
36 |
0.6531 |
0.7027 |
0.4871 |
9.94 |
40 |
0.6405 |
0.7027 |
0.4871 |
10.94 |
44 |
0.6126 |
0.6892 |
0.4871 |
11.94 |
48 |
0.6303 |
0.7027 |
0.4432 |
12.94 |
52 |
0.6264 |
0.7027 |
0.4432 |
13.94 |
56 |
0.6347 |
0.7432 |
0.3669 |
14.94 |
60 |
0.6698 |
0.6622 |
0.3669 |
15.94 |
64 |
0.6346 |
0.7568 |
0.3669 |
16.94 |
68 |
0.6510 |
0.6892 |
0.3704 |
17.94 |
72 |
0.6491 |
0.6892 |
0.3704 |
18.94 |
76 |
0.5947 |
0.7568 |
0.3624 |
19.94 |
80 |
0.6248 |
0.7027 |
0.3624 |
20.94 |
84 |
0.6580 |
0.7027 |
0.3624 |
21.94 |
88 |
0.6345 |
0.7162 |
0.3164 |
22.94 |
92 |
0.6092 |
0.7568 |
0.3164 |
23.94 |
96 |
0.6498 |
0.7162 |
0.2777 |
24.94 |
100 |
0.6915 |
0.7703 |
0.2777 |
25.94 |
104 |
0.6482 |
0.7838 |
0.2777 |
26.94 |
108 |
0.6407 |
0.7973 |
0.2946 |
27.94 |
112 |
0.6135 |
0.7838 |
0.2946 |
28.94 |
116 |
0.6819 |
0.7568 |
0.2546 |
29.94 |
120 |
0.6401 |
0.7568 |
0.2546 |
30.94 |
124 |
0.6370 |
0.7432 |
0.2546 |
31.94 |
128 |
0.6488 |
0.7703 |
0.2477 |
32.94 |
132 |
0.6429 |
0.7973 |
0.2477 |
33.94 |
136 |
0.6540 |
0.7703 |
0.1968 |
34.94 |
140 |
0.5895 |
0.7973 |
0.1968 |
35.94 |
144 |
0.6242 |
0.7568 |
0.1968 |
36.94 |
148 |
0.6575 |
0.7568 |
0.2235 |
37.94 |
152 |
0.6263 |
0.7703 |
0.2235 |
38.94 |
156 |
0.6225 |
0.7838 |
0.2005 |
39.94 |
160 |
0.6731 |
0.7703 |
0.2005 |
40.94 |
164 |
0.6844 |
0.7703 |
0.2005 |
41.94 |
168 |
0.6550 |
0.7703 |
0.2062 |
42.94 |
172 |
0.6700 |
0.7703 |
0.2062 |
43.94 |
176 |
0.6661 |
0.7703 |
0.1933 |
44.94 |
180 |
0.6606 |
0.7838 |
0.1933 |
45.94 |
184 |
0.6757 |
0.7703 |
0.1933 |
46.94 |
188 |
0.6889 |
0.7568 |
0.1895 |
47.94 |
192 |
0.6940 |
0.7568 |
0.1895 |
48.94 |
196 |
0.6919 |
0.7568 |
0.1666 |
49.94 |
200 |
0.6899 |
0.7432 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2
đ License
This model is licensed under the Apache - 2.0 license.
đ Model Index
Property |
Details |
Model Type |
swin-tiny-patch4-window7-224-finetuned-new_dataset_50e |
Training Data |
imagefolder |
Metrics |
accuracy |
Task |
Image Classification |
Evaluation Loss |
0.6407 |
Evaluation Accuracy |
0.7973 |